| 12 Feb 2026 |
 K900 | And I am down to one node still not built and it's paperless | 19:50:01 |
| K900 | (it's always paperless) | 19:50:04 |
| K900 | So I'm going to kick off a new eval and call it a night | 19:51:08 |
 hexa | build time maybe, but runtime is granian | 20:11:28 |
| K900 | It's just the massive closure | 20:34:56 |
| K900 | There's like three opencvs in here | 20:35:06 |
| 13 Feb 2026 |
 Vladimír Čunát | paperless has https://github.com/NixOS/nixpkgs/pull/488795 | 05:27:16 |
|  hoplopf joined the room. | 10:21:21 |
 Marie | is there a reason why previous updates of python3Packages.dbus-fast didn't go through staging?
https://github.com/NixOS/nixpkgs/pull/465361 | 20:37:38 |
| hexa | home-assistant special casing maybe | 20:38:35 |
| Marie | I opened a PR to update this package and chose staging because of the rebuild count, then noticed these previous updates | 20:39:46 |
| hexa | feel free to send into staging | 20:40:00 |
| 14 Feb 2026 |
| K900 | So uh | 08:39:41 |
| K900 | Has anyone touched yarn-berry-fetcher on staging | 08:39:51 |
| K900 | I'm getting hash mismatches on grafana | 08:40:07 |
| K900 | And I'm not getting those on master | 08:40:13 |
| K900 | No changes to the actual code AFAICT | 08:41:03 |
| K900 | This is cursed | 08:41:03 |
 Sergei Zimmerman (xokdvium) | In reply to @k900:0upti.me
No changes to the actual code AFAICT drvPath is the same? | 08:45:43 |
| K900 | No | 09:03:34 |
| K900 | It's staging so a lot different | 09:03:41 |
| K900 | But the actual fetcher code in nixpkgs is the same | 09:03:49 |
 ma27 | if noone beats me, I can bisect this afternoon | 09:14:49 |
 Randy Eckenrode | Redacted or Malformed Event | 11:41:14 |
| Randy Eckenrode | Is anyone able to build Git on any platform on staging? On Darwin, it fails to find a header. On Linux, one of the tests is failing. | 18:41:08 |
 Tristan Ross | I wish I could test that, usually git does fail tests for me but running with less cores helps. We really should split the tests out because it's kinda silly how easily it fails with 128 cores in the test suite. | 20:06:29 |
 alexfmpe | did we become stricter on something output related recently?
pkgsCross.ucrt64.zlib builds on master but not staging-next, where it fails near the end of fixupPhase with
`error: builder for '/nix/store/<blob0>-zlib-x86_64-w64-mingw32-1.3.1.drv' failed to produce output path for output 'static' at '/nix/store/<blob0>-zlib-x86_64-w64-mingw32-1.3.1.drv.chroot/root/nix/store/<blob1>-zlib-x86_64-w64-mingw32-1.3.1-static' | 21:39:52 |
| alexfmpe | * did we become stricter on something output related recently?
pkgsCross.ucrt64.zlib builds on master but not staging-next, where it fails near the end of fixupPhase with
error: builder for '/nix/store/v24848a103agsi7pjm5qj746fi88d5sd-zlib-x86_64-w64-mingw32-1.3.1.drv' failed to produce output path for output 'static' at '/nix/store/v24848a103agsi7pjm5qj746fi88d5sd-zlib-x86_64-w64-mingw32-1.3.1.drv.chroot/root/nix/store/7xlgqiwrifqi4nryhgwvfa9y05myng5b-zlib-x86_64-w64-mingw32-1.3.1-static' | 21:40:23 |
| alexfmpe | and the logs seem to be missing the last shrinking/checking/patching round | 21:41:17 |
 ghpzin | staging-next has:
https://github.com/NixOS/nixpkgs/pull/476830 | 21:42:29 |
